Study design and participants
A cross-sectional community survey was carried out between September 2019 and November 2019 in three poverty communities in Wuhan, Hubei Province, Central China. We chose the poverty communities in three of the seven main regions in Wuhan, namely Hanshui bridge community, Tanhualin community, Zhiyin community, separately located in Qiaokou District, Hanyang District, and Wuchang District. The inclusion criteria were: (a) diagnosed as hypertension, (b) able to communicate in Mandarin, (c) willing to participate in this study, (d) dwelling in low-income communities for more than three months. The exclusion criteria were: (a) difficult in providing answers, (b) having mental disease, (c) unconscious, (d) long-time lying in bed.
The questionnaires were administered by three trained nurses in the community health service center for people or the participants' home. The research purpose, content, significance were well-informed by researchers and informed consents were signed before the investigation. If the participant had blurred vision or unable to write, researchers would read every item and have the answer what they chose. The sample size was decided according to ten times that of the largest item in the questionnaire, and 10% sample loss was considered, totaling at least 132 people. All information were obtained by face to face interviews with trained personnel. Of 326 participants, 31 provided incomplete questionnaire data and the response rate was 90.5%. Finally, a total of 295 were included in our final analysis.
Measurements and instruments
Participant characteristics
A self-designed questionnaire was used to collect information from participates on socio-demographic characteristics (age, gender, marital status, etc) and health-related factors such as course of disease, comorbidity, hospitalization in the past six months, etc. The body mass index (BMI) was calculated as weight (kg) divided by the square of height (m2), according to the Chinese body mass index reference standard[8]. Information on hospitalization was obtained by the responses to the question: “Have you been hospitalized in the past six months? (yes/no) ”.
Measure of HRQOL
We used the SF-12 Questionnaire to measure HRQOL of low-income residents with hypertension. The SF-12, an abbreviated version of the SF-36, has been widely used in the field of HRQOL study[9]. It covers 8 domains with 12 items, including physical function (PF), role limitations due to physical problems (RP), bodily pain (BP), general health (GH), vitality (VT), social function (SF), role limitations due to Emotional problems (RE) and mental health (MH), and measures HRQOL in physical component summary (PCS) and mental component summary (MCS), ranging from 0 to 100[10, 11]. A score more than 50 indicates positive self-rated health[10]. A higher score indicated a better HRQOL and vice versa[12].
Measure of family function
Family function was measured using the Family APGAR Index, which was developed by Smilkstein in 1978[13-15].It is used to assess the satisfaction of family members in five domains: Adaptability, Partnership, Growth, Affection, and Resolve. A 3-point rating scale (0=hardly ever, 1=sometimes, and 2=almost always) was used to score the items. The total score ranged from 0 to 10. A good family function with a score of 7~10, moderate dysfunction of 4~6, and severe dysfunction of 0~3[13-15]. The Chinese version has been widely used with satisfactory validity and reliability[17].
Statistical analysis
Means and standard deviations (SD) were presented for continuous variables, while frequency and percentage were used for categorical variables. We assessed the associations between socio-demographic variables, health-related variables, the Family APGAR index and HRQOL scores using univariate and multivariate analyses. Univariate analyses included a t-tests and one-way ANOVA, whereas multivariate analysis was performed by the clustered multiple linear regression analysis (enter model), where domain scores of the SF-12 instrument were considered as dependent variables and those variables in the three clusters were independent variables. To properly assess the associations between the variables in the 3 clusters and HRQOL, we used dummy variables for disordered multicategory variables.
Specifically, clustered multiple linear regression analyses[18-20]were used to explore the impacts of socio-demographic characteristics, health-related factors, and Family APGAR (3 clusters based on the nature of the study variables and study purpose) on each domain of HRQOL. There was the possibility of multidirectional links among the 3 clusters of independent variables and the dependent variable. In other words, socio-demographic variables (cluster 1) may affect health-related variables (cluster 2) and the Family APGAR (cluster 3) as well as the dependent variables (each domain of HRQOL). Similarly, cluster 2 may affect cluster 3 and the dependent variables. However, cluster 3 may only influence the dependent variables. Consequently, variables in the prior cluster may have impacts on variables in the subsequent cluster, but not vice versa[20].We determined the final regression model in 3 steps, which were described in a previous study[18]: (i) an enter regression of each domain of HRQOL for the cluster 1 variables; (ii) an enter regression for the cluster 2 variables with the equation derived from step 1 as a fixed part of the new regression model; and (iii) an enter regression for the cluster 3 variables, with the equation derived from step 2 as a fixed part of the new regression model. The variables’ inclusion and exclusion criteria for the enter regression models were P values of 0.05 and 0.10, respectively.
The independent effect of each cluster on the dependent variables was determined by calculating the corresponding R2 change. The independent contribution of each cluster was then calculated by (individual R2 change / total R2) ×100%[21].
All statistical analyses were conducted using the Statistical Package for the Social Sciences (SPSS), version 23.0 (SPSS Inc., Chicago, IL, USA). Two-tailed P values below 0.05 were considered statistically significant.